Individual drives cannot reliably be used, in practice, as backup drives. In theory, they can and this is apparently Seagate's position as mentioned below, but as you look at actual backup practice, there are a number of classes of failure which are weaknesses in the single-backup-drive environment. Backup drive hardware failure represents one class of failures. Backup drive firmware failure might represent another, rare, class of failures. OS failures (bugs) and incompatibilities, backup software failures and incompatibilities, and user error, each represent additional classes of possible failure.
One Seagate chat-based tech support apparently believes that a backup drive which has failed is not too serious, because, afterall, it is a BACKUP DRIVE -- meaning the live data is still intact, and it's just that the backup has failed so there is no actual data loss. (Can you believe it?) The possibility that data may have been migrated or that the (inaccessible) backup drive is now being called upon to do a restore, seems not to have occurred to them. (Can you believe it?)
As to backup drive failure classes, of four Maxtor drives we have -- 300 GB, 500GB, 1TB and 1.5TB -- three of them have failed in one way or another. One was an outright hardware failure, one a "false" Maxtor diagnostic error which Seagate tells us to ignore (?), and one a completely inaccessible drive possibly brought about by backup software overrun (1 TB drive out of space!) or maybe the loose thing rattling around inside the case.
In any events, a single external drive unit as backup is in practice a risky thing. The MTBF figures are suspect enough that at least one offsite and one onsite ext drive should be used, paralleling the on- and off-site backup tape storeage used in the old days.
Thanks for that pointer to the article on backup, JA. I take it in the spirit in which it was intended.
Having been involved in data safety and security for our clients and ourselves since 1990, and been a user of Veritas's predecessor, Palindrome in Novell's Netware environment, and an early user of Retrospect on Macs before there was a Windows version, and a user of TAR, BRU, and other backup cmds on SGI's Irix, I still learned a thing or two from the article.
I can't agree with Didrik strongly enough, really. Please understand that the reason that I and other Seagate personnel, especially in Tech Support, are so loud in trumpeting "Have a backup! Have a backup!" is because a VERY large segment of our customers believe that a "backup" consists of having all of one's data stored in an external Seagate/Maxtor/whatever-brand drive and nowhere else. They don't realise that an external drive can fail or lose its partition just as easily as an internal C: drive, and sometimes even more easily since Windows won't let you reformat your C: drive accidentally - you really have to want to!
I personally keep my data in MORE than 2 places. But if you're just a regular user who thinks that your data is 100% safe just because you bought an external drive and cut-and-pasted your data over to it, thereby "freeing up space" (aka erasing it) from its original location, the constant refrain of "Have a backup!" can be a data- and wallet-saver; data recovery is virtually always expensive and often risky, no matter who does it.
So yes, to all who read, please know that even the highest-quality, most expensive hard drive ever produced will fail eventually and it is impossible to predict when that will occur. Given that, it is absolutely essential at all times to keep a backup (better yet, two or more, in different physical locations) of your any data you do not want to lose.